Nehalem, Shanghai, Dunnington SQL性能数据对比(ZT)

肥龙2虽然经过一系列改进，不过由于“落后”的K10架构，同频性能比肥龙1强得有限，略低于肉1，和肉2有一定差距。从国内的测评中，我们只能得出这个结论。如果用肥龙2对比Nehalem，和六核心的Dunnington，结局会怎样？
6 ~3 q' |8 k9 N' m# }6 [4 U电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi会不会让让每个喜欢的AMD的人，产生不忍目睹的感觉呢?( |% W% }" v( ^: l- S, Z$ {

7 Q( O/ a/ Y+ i/ l: J下面的测试，让我们听听真正的龙的咆哮吧。
/ w5 B, p) x! d! a, G9 v交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群
0 l- y0 ~8 R0 Z# O, L* x* F- P华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,PS：对E文没兴趣的朋友，可以直接看这里：华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,' c  T6 B. s2 \7 l& g
测试结果表明，肥龙2在数据库性能上对肉二保持了压倒性的优势，甚至在4核 VS 6核的测试中也毫不吃亏，但是Nehalem的出现将完全扭转这一局面，树立起Intel Xeon对Opteron的性能优势。由于AMD也要推出自己的六核心CPU，到时候六核心Opteron和Nehalem的对比将会更精彩。顺便还有一段6核心（实际只使用了4核）和安腾的性能对比，安腾输得比较凄惨。。。
. K4 i4 Z( n6 K, \/ v/ U$ A玩家论坛
$ g" I2 i% |! `" j2 I3 i7 w6 `; V电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi说明一下：以下为SQL性能方面的测试4 F. M3 X8 e1 ~% N, M& s, d2 j

4 l) S+ b5 L8 A+ p$ X0 J+ _: X7 Y/ h1 F2 }出场选手分别为：玩家论坛6 [: Y" j% c6 X( m% K

1 ?9 L: C% d4 [) M& y1 Z2 X; n+ D玩家论坛Opteron 8384  四核，2M L2，6M L3
; h* \5 i9 [+ C& I, O* F华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,Opteron 8360  四核，2M L2，2M L3bbs.ocer.net3 C, h- \$ V& H9 O9 @
Opteron 2384  四核，2M L2，6M L3交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群1 b( K. b6 w1 ]5 c% B
Xeon X7460    六核，9M L2，16M L3交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群4 y" v: J2 r/ w; h
Xeon X5570    四核，1M L2、8M  L3（Nehalem）
$ g, @) r+ ~, f( y  |4 N电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATiXeon X5470    四核，12M L2
; Q1 z( j. Y' J. S0 s交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群玩家论坛( F5 F" g; q. n% v% I
Nehalem, Shanghai, and Dunnington performance notes
9 j( g( I8 Y2 o& N7 Y华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,
2 U3 x' J: y8 l7 g1 o玩家论坛Back in March 2008, I discussed the SAP SD 2 Tier results for some AMD Opteron and Intel Xeon systems. Here are some more recent results. All Opteron and Xeon processors below are on the 45nm process except for the Opteron 8360, which is a 65nm product. For some reason, HP has not posted a 4-socket result for the 45nm Opteron 8384, where the bigger 6M L3 cache in known to improve network round-trip performance. The Itanium is a dual-core 90nm product, which is at serious disadvantage (to be replaced Tukwila, a 65nm quad-core). Note that the Xeon 7460 is a six core processor. My expectation is that the 4-way Opteron 8384 should be around 22-23K which would be very competitive for a quad-core going against a six-core.
# M; T. C) l, r" G2 X交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群) ?! n# w7 c0 i1 D2 t

' F0 D5 h1 }3 g6 k* I6 G华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,) f0 G! G. c+ R2 u9 h7 v
System             Processors                                           Users                SAPSbbs.ocer.net* x5 g2 f6 J  {0 }" r0 ~

, [$ i' j' F5 ?0 ?6 w交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群DL785G5           8 x 2.7GHz        Opteron 8384                7,101                 35,400 (O10,Lin)bbs.ocer.net+ P8 n7 F( A: n0 H4 W& ~6 Q
交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群- b( _* e  Q- K/ D/ W
DL580G5           4 x 2.66GHz      Xeon X7460                   5,155                 25,830 (S2K5)
1 F( k' ]! b- x华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,
# L/ S4 U! M, X; c玩家论坛DL380G6           2 x 2.93GHz      Xeon X5570                   4,995                 25,000 (S2K5)
* t$ Q$ S2 Q. R9 t* H% m& j: ]bbs.ocer.net
4 ^# c: F" I  c- P2 bbbs.ocer.netDL585G5           4 x 2.5GHz        Opteron 8360                3,801                 19,020 (S2K5)bbs.ocer.net" g. w$ X$ W8 L3 V. Z

2 y  [. S4 }% N华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,DL385G5           2 x 2.7GHz        Opteron 2384                2,752                 13,780 (S2K5)
, F. O( g+ A' g  j- z0 Wbbs.ocer.net
8 T* B# d6 |5 X) E交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群DL380G5           2 x 3.33GHz      Xeon 5470                     2,518                 12,600 (S2K5)bbs.ocer.net1 j* l, Z( t& h, G! l0 t

) |& K+ d2 m8 e* z华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,BL860C             2 x 1.66GHz      Itanium 9140M                  501                  5,850华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,( M( C+ v' @' {3 L

) `+ b. T0 ]0 F: u  J3 f4 k& ], n华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,
! P" ^1 L. f# @( c( u, a华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,bbs.ocer.net6 |9 a- K9 _+ L1 Y2 A
Of special note is the exceptional result for the 2-socket Xeon X5570, based on the new Nehalem processor, expected to be available some time in the first half of 2009. This is not entirely unexpected. First, the integrated memory controller probably helps in network round-trip intensive operations. Second is the return of Hyper-Threading, as Nehalem was designed in Oregon, while the Core 2 architecture is an Israeli design (each design team has their own opinion of various micro-architecture features). On the last Oregon design, the Pentium 4 NetBurst core, I noted that Hyper-Threading improved network round-trip performance by 15-20%, but did not actually improve any SQL operation outside of the network round-trip. Later, I measured 40-50% performance gains for HT on LiteSpeed backup compression tests. I have heard that while the theory behind HT is sound, certain operations such as acquiring locks (at the C/C++ level) can cause problems. The compression algorithm has no such issues, and probably represents the upper bound on what could be achieved with HT. Supposedly the Itanium HT, which was introduced with Montecito after the last NetBurst, had some improvements over NetBurst. Now that the Oregon team has had 8+ years to investigate HT characteristics, we should expect a much improved HT with Nehalem.电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi$ {; I* M+ i" y/ a5 u

3 Y6 F/ t7 v- e4 p华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频, 交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群- S: o4 A) C4 F( D! y

3 k" F! P. j" r( N" ]  G电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATiMy expectation is HT has the biggest benefit in SAP type environments, i.e. stored procedure calls that retrieve a single row (50-100 CPU-micro-sec), moderate benefit in TPC-C and E type environments (on the order of 1-3 CPU-milli-sec), and less or no benefit in large TPC-H type queries. When I get the Nehalem Xeon 55xx system, I will look into this.
2 Z6 q+ U, e( w9 j" C4 B华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,
- P7 \. j4 W# Y玩家论坛 玩家论坛* e3 @9 n( x( A5 z2 G

  Y" y" ~' k" o: f1 M0 V交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群Unisys to focus on Xeon over Itanium?
' M: j* M" n' g3 a2 d( F交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群0 S! [* K3 q6 W$ {
On a side note, see the article regarding Unisys.
  U: k: h+ a6 A% t: Y( W( X  ^交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群/ \0 [9 q$ R- C. v  U: ]
http://news.cnet.com/8301-13924_ ... s&tag=2547-1_3-0-20玩家论坛9 N/ ~9 G( J7 ^3 a* K

8 e& [( W. N# l1 {$ ]0 lUnisys just posted a 10TB TPC-H result for a 16-socket Xeon X7460. While the X7460 has six cores, Unisys only enabled four cores per socket, in part, because the current version of Windows and SQL Server only support up to 64 cores. The result is 26% higher than for a 32-socket 64 core Itanium 2 system. The presumption is that later this year, Intel will release the quad-core Itanium, codename Tukwila, so this result might be representative of 16-socket systems in late 2009. Even if Tukwila can achieve 2.0GHz, it will probably just be comparable to the X7460. After Windows Server 2008 R2 releases, the 16-way X7460 will have all 72 cores available for the through-put portion of TPC-H.玩家论坛; f4 Q6 }& `3 J: f8 b
电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi0 [& e4 I+ J+ b

" [  {0 _$ c  R% B) c玩家论坛电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi5 P3 N3 U8 i7 g+ K# B
64-core TPC-H 10TB
  B5 G5 e0 s* n& M9 p1 l
5 F9 j5 b% }- ~( {, A) J玩家论坛Xeon X7460 Six-core 2.66GHz (Dunnington, 4 cores used) 80,172.7 (ES7600R)玩家论坛& N* c3 p7 k0 b6 p

$ ~2 `4 p; D4 R/ a9 b: m% M: e华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,Itanium 9140N Dual-core 1.6GHz (Montecito) 63,650.9 (Superdome)
2 E/ I4 J7 o* j# s5 d华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,' M- f" n1 L1 t" C. s" X

* g# m, F9 _- g, v- X0 A7 `电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi华硕,金士顿,金邦,技嘉,悍马,悍丝俱乐部,映泰,极限超频,超频,% A4 A- ]4 B  O) Q, u0 a# q- x
Dell TPC-E results for Shanghai versus Dunnington电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi8 g* h. J8 C, l# A9 s" n8 G
电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi/ g8 F. W2 X2 o3 e- Y
Both systems 4 sockets, 64GB memory,
. A/ O' Q, u# p4 N电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATibbs.ocer.net5 B3 h( U# i! {3 s8 o
4 x six-core Dunnington 2.66GHz, 16M L3   671.35 tpsE
* x5 F! o' q  m* N0 b2 R- ^2 ~玩家论坛交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群! ]; m  N: L  f2 S5 V1 T. S
4 x quad-core Shanghai 2.7GHz, 6M L3      635.43 tpsE交互净土◆玩家天堂——致力于打造中国最大的电脑用户论坛集群$ Y4 S3 j/ E9 N( |' t  I
玩家论坛7 J8 ]& g, A( r# a2 q' Q+ A  f
So Shanghai quad-core competes very well against Dunnington with six cores. The large cache relative to Barcelona (2M L3) really helped
( t3 M0 `2 ?8 t7 I2 a5 \! H电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi电脑|发烧|玩家|超频|改造|Intel|AMD|nVidia|ATi, {% l' o$ o# C; f7 Z
[ 本帖最后由 chaosXP 于 2009-3-25 11:59 编辑 ]